16F88 PWM and A/D conversion

[williB]

They even stack the zeners, when charging the supercaps ..
so far i got a very small program , programmed and running on the 628..yay !! ya gotta remember its been months since i programmed a pic..
next step is to program the 16F88..i understand that the A/D converter is selected and available from the get go..?
this should be fun.. 8)

 

[Nigel Goodwin]

They even stack the zeners, when charging the supercaps ..

I thought the idea was to charge the capacitors?, the zeners will probably absorb more energy than the capacitors get :lol:

 

[williB]

i'll link you their diagram , in a second..
but check out their product line..!!
**broken link removed**
they can be found on
www.maxwell.com

 

[williB]

i cant find it .. but what i saw showed several layers of caps in parallel with zener diodes , to make sure that each caps voltage did not excede the 5.5 V maximum..

 

[williB]

I got them ..
and tried two in parallel , with a NiMH cell below them ..
so i charged caps up to 5V then connected the cell , then connected that to a 100 ohms resistor , then connected diode ..
diode current started @ 13 mA and 4.5 mins. later current was 1.3 mA..
so it should last twenty mins. without blinking at all . but i still want to use PWM and the A/D converter of the PIC 16F88..

 

[williB]

sadly i'm not getting anywhere with my PWM routine
i need help.. getting started..
i have been looking at the PDF for days !! :cry:
if i have a templet to go by , i can make changes as needed..
What i need is a 10 precent duty cycle PWM signal running at 100 HZ with a 5.068 Mhz clock on a 16F628..
thanks in advance

 

[upand_at_them]

Have you tried Nigel's tutorial #8?

Mike

 

[williB]

yeah i've seen it

 

[williB]

The next step is to set T2CON, this sets the frequency of the PWM, as it's derived from the 20MHz system clock it runs too at too high a frequency, there are two possibilities here - setting the prescaler divides the frequency before the PWM section, and the postscaler afterwards. For this example we set the prescaler to divide by 16, this gives us a PWM frequency of 2500Hz.

is he saying that 20MHZ /16 = 2500HZ ?

 

[williB]

thanks for all the help :wink:

 

[motion]

The next step is to set T2CON, this sets the frequency of the PWM, as it's derived from the 20MHz system clock it runs too at too high a frequency, there are two possibilities here - setting the prescaler divides the frequency before the PWM section, and the postscaler afterwards. For this example we set the prescaler to divide by 16, this gives us a PWM frequency of 2500Hz.

is he saying that 20MHZ /16 = 2500HZ ?

The calculation should be as follows:

PWM Period = [(PR2) + 1] • 4 • Tosc • Prescale (see section 8.3.1 of the PIC16F87XA datasheet)
PWM Period = [124+1] * 4 * 0.05usec * 16
PWM Period = 400usec. for a 2500Hz frequency.

The correct PR2 value should be 124 not 126 to get 2500Hz.

 

[williB]

wow thanks !

 

[williB]

so if my clock is 39600hz
prescale is 1
pwm period is .001sec (1000 HZ)
then
PR2 = 9.. right?

 

[motion]

so if my clock is 39600hz
prescale is 1
pwm period is .001sec (1000 HZ)
then
PR2 = 9.. right?

Yes. PR2 is about equal to 9. However, if you set the PR2 to a very low value, the number of PWM duty cycle slices steps will be very low. In this case, it's just 36 slices (9*4).

 

[williB]

Great !!
so my clock is 5,068,800 HZ divide that by 8..gives 633,600 Hz clock into PIC..
with a 1000 HZ PWM Freq.
and a prescale of 1..
gives
.001 = ((PR2+1) * 4) / 633600
PR2= 157
and my number of different steps is 633 .
that really helped , because i saw that equation but without your explaination it didnt sink in..

 

[Nigel Goodwin]

Great !!
so my clock is 5,068,800 HZ divide that by 8..gives 633,600 Hz clock into PIC..
with a 1000 HZ PWM Freq.
and a prescale of 1..
gives
.001 = ((PR2+1) * 4) / 633600
PR2= 157
and my number of different steps is 633 .
that really helped , because i saw that equation but without your explaination it didnt sink in..

I suggest you try reading my tutorial, which makes thinks a little clearer - PR2 basically sets the number of steps, but the tutorial gives the simple sequence of steps that I used.

 

[williB]

i think i finally have it.. Yay !
*goes off to test it out*

 

[williB]

isnt the addres of TrisB on a 628 86h ??
i think that page was wrong ..

 

[motion]

Great !!
so my clock is 5,068,800 HZ divide that by 8..gives 633,600 Hz clock into PIC..
with a 1000 HZ PWM Freq.
and a prescale of 1..
gives
.001 = ((PR2+1) * 4) / 633600
PR2= 157
and my number of different steps is 633 .
that really helped , because i saw that equation but without your explaination it didnt sink in..

I suggest you try reading my tutorial, which makes thinks a little clearer - PR2 basically sets the number of steps, but the tutorial gives the simple sequence of steps that I used.

Tutorial8 needs correcting. The PWM frequency calculations don't add up. That caused WillieB to ask for help.

 

[williB]

https://www.electro-tech-online.com/custompdfs/2005/06/33023a.pdf
page 216
i think this is wrong also

CLRF CCP1CON ; CCP Module is off
CLRF TMR2 ; Clear Timer2
MOVLW 0x7F ;
MOVWF PR2 ;
MOVLW 0x1F ;
MOVWF CCPR1L ; Duty Cycle is 25% of PWM Period
CLRF INTCON ; Disable interrupts and clear T0IF
BSF STATUS, RP0 ; Bank1
BCF TRISC, PWM1 ; Make pin output
CLRF PIE1 ; Disable peripheral interrupts
BCF STATUS, RP0 ; Bank0
CLRF PIR1 ; Clear peripheral interrupts Flags
MOVLW 0x2C ; PWM mode, 2 LSbs of Duty cycle = 10
MOVWF CCP1CON ;
BSF T2CON, TMR2ON ; Timer2 starts to increment
;
; The CCP1 interrupt is disabled,
; do polling on the TMR2 Interrupt flag bit
;
PWM_Period_Match
BTFSS PIR1, TMR2IF
GOTO PWM_Period_Match
;
; Update this PWM period and the following PWM Duty cycle
;
BCF PIR1, TMR2IF

fifth line down MOVLW 0x1F ;
is indeed 25% of the PWM Period

thirteenth line down MOVLW 0x2C ; PWM mode, 2 LSbs of Duty cycle = 10
but when you tack on the two LSBs of the duty cycle
it dosnt make any sence..